Research project portfolio

Clinical problem

Patients and clinicians increasingly hold molecular reports that are technically complete but difficult to interpret. The barrier is not access to the result; it is understanding what the result means, how confident it is, and what question it raises next.

Scientific question

Can molecular findings be explained in accurate, appropriately hedged language that preserves scientific uncertainty rather than flattening it?

Data

Molecular variant descriptions, curated variant knowledge bases and published biomedical literature.

Methods

Natural language processing over biomedical text, structured variant annotation, and retrieval grounded in cited sources rather than unsourced generation.

Innovation

The design goal is explanation with its uncertainty intact — including saying plainly when the evidence for a finding is weak or absent.

Potential clinical impact

Better-prepared conversations between patients and their treating teams, and a clearer view of which findings warrant discussion.

Outputs

In development.

Clinical problem

Molecular tumour board discussions vary widely in structure, and the reasoning behind a recommendation is often not recorded in a form that can be revisited.

Scientific question

What minimum structure allows a molecular tumour board decision to be recorded so that its evidence basis and its uncertainties remain legible afterwards?

Data

De-identified case structure, evidence tiers and actionability categories.

Methods

Evidence tiering frameworks, structured decision summaries, and standardised categorisation of uncertainty.

Innovation

Recording what was not concluded, and why, alongside what was — so that a later reviewer can see where the reasoning was thin.

Potential clinical impact

More consistent molecular decision-making and a record that supports audit and improvement over time.

Outputs

In development.